DRAFT VERSION OCTOBER 2, 2018 Typeset using LATEX twocolumn style in AASTeX61 ULTRA-WIDEBAND DETECTION OF 22 COHERENT RADIO BURSTS ON M DWARFS ∗ JACKIE VILLADSEN1, 2 , AND GREGG HALLINAN2 1National Radio Astronomy Observatory 520 Edgemont Rd Charlottesville, VA 22903 2California Institute of Technology 1200 E California Blvd, MC 249-17 Pasadena, CA 91125 ABSTRACT Coherent radio bursts detected from M dwarfs have some analogy with solar radio bursts, but reach orders of magnitude higher luminosities. These events trace particle acceleration, powered by magnetic reconnection, shock fronts (such as formed by coronal mass ejections, CMEs), and magnetospheric currents, in some cases offering the only window into these processes in stellar atmospheres. We conducted a 58-hour, ultra-wideband survey for coherent radio bursts on 5 active M dwarfs. We used the Karl G. Jansky Very Large Array (VLA) to observe simultaneously in three frequency bands covering a subset of 224- 482 MHz and 1-6 GHz, achieving the widest fractional bandwidth to date for any observations of stellar radio bursts. We detected 22 bursts across 13 epochs, providing the first large sample of wideband dynamic spectra of stellar coherent radio bursts. The observed bursts have diverse morphology, with durations ranging from seconds to hours, but all share strong (40-100%) circular polarization. No events resemble solar Type II bursts (often associated with CMEs), but we cannot rule out the occurrence of radio-quiet stellar CMEs. The hours-long bursts are all polarized in the sense of the x-mode of the star’s large-scale magnetic field, suggesting they are cyclotron maser emission from electrons accelerated in the large-scale field, analogous to auroral processes on ultracool dwarfs. The duty cycle of luminous coherent bursts peaks at 25% at 1-1.4 GHz, declining at lower and higher frequencies, indicating source regions in the low corona. At these frequencies, active M dwarfs should be the most common galactic transient source. Keywords: stars: flare — stars: coronae — radio continuum: stars arXiv:1810.00855v1 [astro-ph.SR] 1 Oct 2018 Corresponding author: Jackie Villadsen [email protected], [email protected] ∗ Jackie Villadsen is a Jansky Fellow of the National Radio Astronomy Observatory. 2 VILLADSEN ET AL. 1. INTRODUCTION ations of exoplanet habitability (Shields et al. 2016). In con- The Sun produces intense bursts of emission at low ra- trast to solar-type stars, M dwarfs retain their youthful rapid dio frequencies, which caused it to become the second de- rotation (Newton et al. 2016) and vigorous magnetic activ- tected astrophysical radio source after the Galactic center ity (West et al. 2008) for hundreds of millions to billions of (Hey 1946; Southworth 1945; Orchiston 2005). These bursts years. Even old, slowly rotating M dwarfs can have signifi- are classified according to their morphology in the time- cant magnetic spot coverage (Newton et al. 2016) and pow- frequency plane (“dynamic spectrum”), reviewed in Dulk erful flares (France et al. 2016). Active M dwarfs have a high (1985) and Bastian et al.(1998). There are a large num- rate of energetic flares (Lacy et al. 1976), with potentially ber of types of solar radio bursts, with durations from mil- dramatic consequences for planetary atmospheres: coronal liseconds to weeks, at frequencies from kHz to GHz. Many mass ejections can compress the planetary magnetosphere of the bursts have high brightness temperatures, strong cir- and erode the atmosphere (Khodachenko et al. 2007; Lam- cular polarization, and/or narrow bandwidth, indicating that mer et al. 2007; Kay et al. 2016; Patsourakos & Georgoulis they are powered by coherent emission mechanisms, plasma 2017) and stellar energetic particles can penetrate deep into emission or electron cyclotron maser (ECM) emission. As a the atmosphere and interact chemically (Segura et al. 2010; picture has emerged for the origin of the different types of Airapetian et al. 2016; Tilley et al. 2017; Howard et al. 2018). bursts, they have proved to have significant diagnostic power There is limited observational evidence of CMEs and ener- for processes in the solar corona and interplanetary medium, getic protons around stars other than the Sun, so models of including Earth-impacting (“geoeffective”) events. the impact of stellar ejecta on exoplanetary atmospheres must Two tracers of geoeffective space weather events are solar rely on extrapolating solar relationships, between flares and Type II and Type III bursts, which occur at frequencies from CMEs (Yashiro & Gopalswamy 2009; Aarnio et al. 2011; hundreds of MHz down to kHz. These bursts occur at the Drake et al. 2013; Osten & Wolk 2015; Odert et al. 2017), local plasma frequency and its second harmonic, and their magnetic flux and CMEs (Cranmer 2017), and flares and en- frequency declines over time as a coronal source moves out- ergetic protons (Belov et al. 2007; Cliver et al. 2012; Atri wards into lower densities. Solar Type II bursts trace coronal 2017; Youngblood et al. 2017), to apply to the high-energy shock fronts, moving at speeds of order 1000 km s-1, passing flares and strongly magnetized coronae of active M dwarfs. through a decade in plasma frequency in a matter of min- There are both observational and theoretical reasons to ex- utes. These shocks may be driven by coronal mass ejections pect that solar scaling relationships for CMEs and energetic (CMEs) or by flare blast waves (Vršnak & Cliver 2008). So- protons may not apply to active M dwarfs. Observations of lar Type III bursts are produced by electron beams moving Lyman α absorption in the astrosphere of active M dwarf at speeds of order 10% of the speed of light, with radio fre- EV Lac indicate that it has a mass loss rate similar to the Sun quency decreasing exponentially on a timescale of seconds; (assuming isotropic mass loss; Wood et al. 2005), in contrast when these bursts occur at low frequencies, they indicate that to the flare-based prediction of Osten & Wolk(2015) that it energetic electrons (and likely protons as well) are escap- would be up to 500 times greater due to frequent CMEs. The ing the solar corona on open magnetic field lines. Both of strong global field of active M dwarfs (kG strength - Morin these classes of bursts are correlated with solar space weather et al. 2008) may confine or prevent eruptions (Vidotto et al. events that impact Earth: coronal mass ejections and solar 2016; Odert et al. 2017; Alvarado-Gómez et al. 2018), as energetic particle (SEP) events (Kouloumvakos et al. 2015). has been suggested for the few strong solar flares that are Analogous events, tracing bulk motion of plasma, may be de- not accompanied by CMEs (Thalmann et al. 2015; Sun et al. tectable from nearby stars; and based on Sun-as-a-star data, 2015). In addition, the most intense solar energetic parti- such stellar radio bursts can be used to recover approximate cle events that hit Earth are caused by particle acceleration dynamical properties of CMEs (Crosley et al. 2017). This at CME shocks (Reames 2013), so solar flare-proton scaling suggests that low-frequency radio observations can be used relationships will not apply well to active stars if energetic to characterize stellar mass loss and the space weather envi- flares are not accompanied by CMEs or if those CMEs do ronment experienced by extrasolar planets. not form shocks to accelerate particles; shocks may be rare Based on planet occurrence statistics, many of the near- due to high Alfvén speeds in the strongly magnetized stel- est habitable-zone Earth-size planets should be found around lar coronae. With all these complicating factors, there is a M dwarfs (Dressing & Charbonneau 2015), a prediction need for direct stellar observations to test for the presence of borne out by the planet detections around Proxima Cen CMEs, shocks, and energetic particles. (Anglada-Escudé et al. 2016), TRAPPIST-1 (Gillon et al. Beyond radio observations, various other observational 2017), and Ross 128 (Bonfils et al. 2017). However, the techniques provide evidence suggestive of stellar erup- space weather environment around M dwarfs may differ dra- tions. Chromospheric and coronal emission lines some- matically from our own solar system, complicating consider- times display blueshifted emission components during flares COHERENT RADIO BURSTS ON MDWARFS 3 on M dwarfs (e.g., Houdebine et al. 1990; Fuhrmeister & II and III radio bursts) have been inconclusive: while early Schmitt 2004; Leitzinger et al. 2011; Vida et al. 2016). Sur- observations yielded bright events coincident with optical veys with many hours of stellar observations (Leitzinger flares (e.g., Lovell 1963), these events may have been due et al. 2014; Vida et al. 2016) have detected these blueshifted to RFI. Later experiments imposed additional requirements features at a rate much lower than the rate of energetic flares, to distinguish between stellar bursts and RFI, including: in- but Leitzinger et al.(2014) attribute this low rate of CME terferometric observations (finding two hours-long 408-MHz detections to sensitivity limits. bursts on YZ CMi; Davis et al. 1978), absence of the sig- Harra et al.(2016) searched for ways of differentiating be- nal in off-source beams (identifying a handful of candidate tween eruptive and non-eruptive energetic solar flares, using events; e.g., Abdul-Aziz et al. 1995; Abranin et al. 1998), optical, EUV, and X-ray signatures potentially observable in and spectral shape of the burst (yielding a short 430-MHz stars. They ruled out long flare decay timescales (suggested burst on YZ CMi; Bastian et al. 1990). Other low-frequency as evidence of a stellar eruption on AU Mic; Cully et al. 1994; observations produced marginal or uncertain detections (e.g., Katsova et al. 1999) and identified only one reliable indicator Jackson et al.